In general, devices for detecting and generating a warning with respect to dangerous conditions, such as the presence of combustion products or carbon monoxide, are known. For example, various smoke detector systems are described in U.S. Pat. No. RE 33,920, reissued on May 12, 1992, to Tanguay et al; U.S. Pat. No. 4,870,395 issued Sep. 26, 1989, to Belano; and U.S. Pat. No. 4,965,556 issued Oct. 23, 1992, to Brodecki et al, all the foregoing referenced patents being commonly assigned with the present invention. Other examples of such detectors are described in U.S. Pat. No. 3,932,850 issued to Conforti et al on Jan. 13, 1976; U.S. Pat. No. 4,020,479 issued to Conforti et al on Apr. 26, 1977; U.S. Pat. No. 4,091,363 issued to Siegel et al on May 23, 1978; U.S. Pat. No. 4,097,851 issued to Klein on Jun. 27, 1978; U.S. Pat. No. 4,225,860 issued to Conforti Sep. 30, 1980; U.S. Pat. No. 4,258,261 issued to Conforti on Mar. 24, 1981; U.S. Pat. No. 4,302,753 issued to Tice on Aug. 8, 1995; U.S. Pat. No. 5,473,167 issued to Minnis on Dec. 5, 1995; U.S. Pat. No. 5,483,222 issued to Tice on Jan. 9, 1996; U.S. Pat. No. 4,097,851 issued to Klein on Jun. 27, 1978; U.S. Pat. No. 4,138,664 issued to Conforti on Feb. 6, 1979; U.S. Pat. No. 4,138,670 issued to Schneider et al on Feb. 6, 1979; U.S. Pat. No. 4,139,846 issued to Conforti on Feb. 13, 1979; U.S. Pat. No. 4,225,860 issued to Conforti on Sep. 30, 1980; U.S. Pat. No. 4,287,517 issued to Nagel on Sep. 1, 1981; U.S. Pat. No. 4,829,283 issued to Spang et al on May 9, 1989; U.S. Pat. No. 5,172,096 issued to Tice et al on Dec. 15, 1992; U.S. pat. No. 5,422,629 issued to Minnis on Jun. 6, 1995; and U.S. Pat. No. 5,440,293 issued to Tice on Aug. 8, 1995.
Most combustion product detectors employ ionization chamber and/or photoelectric sensors. Carbon monoxide (CO) detectors are also known. In general, CO detectors employ one of three types of detectors: semiconductor, biomimetic and electrochemical.
Semiconductor CO sensors typically employ a thin layer of metal, such as tin dioxide, maintained at a relatively high temperature (e.g., 100.degree. C. to 400.degree. C.). The surface conductivity of the metal varies generally proportionally in accordance with exposure to ambient CO concentration. The semiconductor chip measures the migration of oxygen molecules through the surface of the sensor material. Such semiconductor CO sensors have drawbacks in that they have relatively high power requirements and are therefore not practical for battery units. In addition, many semiconductor CO sensors require high temperature (e.g., 400.degree. C.) purging to burn off attracted CO on a periodic basis; e.g., every 2.5 minutes. There is also difficulty in determining the efficiency or working condition of semiconductor CO sensors; self-diagnostic tests are not generally available. In addition, semiconductor CO sensors tend to be sensitive to other gases in addition to carbon monoxide, giving rise to a potential for false alarms, and sensor accuracy can drift substantially (up to 40%) over time.
Biomimetic sensors utilize a transparent substrate disk coated with a synthetic hemoglobin that mimics the reaction of natural hemoglobin in the presence of carbon monoxide. The biomimetic material darkens with cumulative absorption of CO. A light emitting diode (LED) transmits light through the biomimetic material to a photosensitive device. When the material becomes sufficiently dark to prevent adequate light from reaching the photosensitive device, the detector sounds an alarm. An example of a biomimetic sensor is described in U. S. Pat. No. 5,063,164 issued to Goldstein on Nov. 5, 1991.
Biomimetic sensor based systems are disadvantageous in a number of respects. The time period necessary for the sensor to recover from exposure to carbon monoxide is relatively long time (e.g., 24 to 48 hours). Thus, assuming that the alarm system is silenced until the sensor recovers, the occupants of the home are unprotected during that period. In addition, exposure to particularly high levels of CO can permanently darken the sensor. Further, biomimetic sensors are susceptible to generating false alarms because their self-diagnostic capabilities tend to be limited.
Electrochemical sensors, in general, employ a chemical reaction to convert CO to carbon dioxide (CO.sub.2) to create a chemical imbalance in a portion of the cell which in turn generates a current indicative of the amount of CO present. Some electrochemical sensors utilize two chambers (one for CO and one for hydrogen). However, calibration of the sensor is required, and self-diagnostic capabilities tend to be limited.
Various standards have been set with respect to the performance of dangerous condition alarms for residential use. For example, Underwriters Laboratory (UL) in the United States and Canada have promulgated standards UL 217, ULC-S531, UL 268 and ULC-S529 with respect to smoke detectors and UL 2034 (effective Oct. 1, 1995) with respect to CO detectors.
UL standards for dangerous condition alarm systems for residential use typically define certain alarm conditions. For example, UL 2034, requires that a CO detector generate an alarm in response to cumulative exposure to CO concentrations at specified levels measured in parts per million (PPM) within predetermined time periods (e.g., sound an alarm at 100 PPM in less than 90 minutes, 200 PPM in less than 35 minutes and 400 PPM in less than 15 minutes). However, in order to reduce nuisance alarms, the UL standard also requires that a CO detector ignore cumulative exposure to various low concentrations of CO for minimum time periods (e.g. 15 PPM for up to 30 days, with additional exposure to 35 PPM for one hour twice a day to simulate potential cyclical changes in CO levels resulting from vehicle traffic, 60 PPM for up to 28 minutes, and 100 PPM for up to 16 minutes).
In addition, UL standards sometimes require that dangerous condition alarms incorporate some manner of manually actuable reset button. For example, UL 2034 requires that a CO detector include a manually actuable reset button which, in effect, decreases the sensitivity of the device and turns off the alarm for a predetermined time period. If the CO concentration is maintained or continues to rise at the conclusion of the reset period (defined by UL 2034 as being a maximum of six minutes), then the alarm will be re-actuated.
UL standards often also require that dangerous condition alarm devices be marked with specific warning and/or operating instructions. For example, UL 2034 requires that a CO detector be marked with certain operating instructions which set forth a particular protocol to be followed in the event that the alarm sounds. The instructions advise the occupant to call the fire department only if someone is experiencing symptoms of CO poisoning (headache, dizziness, upset stomach, etc.). If no CO poisoning symptoms are present, the occupant is instructed to reset (silence) the detector and investigate the source of the CO.
Given the nature of the dangers protected against by such dangerous condition warning devices, it is particularly important that the sensors be reliable and relatively foolproof. This need is accentuated when the unit employs a DC power source and/or replaceable sensor unit. It is therefore important to ensure that replaceable units be installed properly, (e.g., are not reversed during installation), are in good operating condition, and that an occupant be given sufficient warning of an impending sensor or battery failure. In general, generation of a low battery warning signal is known. Examples of apparatus for generating an alarm to indicate impending battery failure in the context of a battery powered fire detector are described in U.S. Pat. No. 4,139,846 issued to Conforti on Feb. 13, 1979; U.S. Pat. No. 4,138,670 issued to Schneider et al on Feb. 6, 1979; and U.S. Pat. No. 4,138,664 issued to Conforti et al on Feb. 6, 1979.
Another source of frustration with dangerous condition detectors is the inability of the typical user to discern which of a number of detector units is generating warning signals as to impending battery or sensor failure. Conventionally, a low battery warning signal is generated by intermittent actuation of the same horn used to generate a danger condition alarm. The low battery warning signal is distinguishable from a danger condition alarm by the duty cycle and/or repetition rate. However, it is often very difficult to localize sound. This difficulty tends to be exacerbated when the units are mounted in inaccessible places such as, for example, on a cathedral ceiling, or are mounted in close proximity to other devices, such as other dangerous condition detectors (e.g., a CO detector mounted near a smoke alarm). Some detectors also include a visual indicator, such as an LED, that blinks in synchronization with the low battery audible alarm, albeit not coincidentally. However, to conserve battery power, the LED activation is held to a relatively short duration, e.g., 10 milliseconds, and the repetition rate is typically kept relatively low, e.g., one flash each 40 seconds. As a result, unless the user happens to be looking in the direction of the unit when the LED flashes or is able to correlate a 10 millisecond flash with a 10 millisecond chirp delayed by several seconds, it is difficult to identify the particular unit in distress.
It is desirable to provide a brief respite, as through a night, from a distinctive, repeating audio alarm if the battery of a battery operable dangerous condition sensor is determined to be failing, but failure is not imminent; however, if battery failure is imminent, it would be unwise to ignore the consequent alarm. It is to a feature for accommodating these contrasting conditions that the present invention is directed.